Quality prognostics system and method for manufacturing processes

Abstract

A quality prognostics system and a quality prognostics method for predicting the product quality during manufacturing processes are disclosed. The present invention utilizes the current production tool parameters sensed during the manufacturing process and several previous quality data collected from the measurement tool to predict the future product quality. The quality prognostics system is composed of conjecture modeling means and prediction modeling means. The conjecture modeling means itself also can be applied for the purpose of virtual metrology. Further, the quality prognostics method possesses a self-searching means and a self-adjusting means for searching the best combination of various parameters/functions used by the conjecture algorithm or prediction algorithm; and meeting the requirements of new equipment parameters and conjecture/prediction accuracy.

Description

RELATED APPLICATIONS [0001] The present application is based on, and claims priority from, Taiwan Application Serial Number 93115993, filed Jun. 3, 2004, the disclosure of which is hereby incorporated by reference herein in its entirety. FIELD OF THE INVENTION [0002] The present invention relates to a quality prognostics system and a quality prognostics method for manufacturing processes, and more particularly, to the quality prognostics system and method suitable for use in virtual metrology and predicting product quality manufactured on production tools in semiconductor or TFT-LCD manufacturing plants. BACKGROUND OF THE INVENTION [0003] Production and quality measurement are generally separated in semiconductor and TFT-LCD plants, i.e. after first being processed on a production tool, products are then delivered to a measurement tool for inspection. Based on cost consideration, most of the inspection jobs are done by randomly selecting some products in a product lot to determine t...

AI Technical Summary

Benefits of technology

[0017] Hence, with the application of the present invention, the quality of a next product lot can be prognosed before being produced by using the current process parameters of a production tool and the actual measurement values of several previous product lots produced in the measurement tool; and the present invention also has generic applicability, thus reducing the shortcomings of the conventional skills.

Problems solved by technology

Consequently, occurrence of defects during production process may not usually be discovered until measurement.

As such, numerous defective products might have been produced before measurement is performed.

Those methods are commonly disadvantageous in that defective products have already been produced as soon as the monitoring system detects that the process parameters are abnormal.

Therefore, defects not only reduce product yield, but also seriously impact production capacity and cost.

However, the above studies mainly focus on using the available sensor data to identify possible defects of current production, but not the quality of the next product lot.

Generally speaking, not many input data can be considered simultaneously for purposes of conjecture in most studies, and the applications of the proposed conjecture schemes are limited to certain types of equipment.

This patent reference cannot conjecture the film thickness by the method of virtual metrology, and the film thickness cannot be determined until measurement.

This method is disadvantageous in that: varieties and difficulty level of the data-based model relatively increase a lot, when there are many process parameters; and the data-based model cannot be properly used in accordance with various tool features, thus not having flexibility.

However, this method is merely based on the existing parameter data to predict the measurement value of the product currently being manufactured, but cannot further predict the quality of the product to be produced in a future period of time.

However, the method disclosed in this patent reference basically can be merely applicable to certain types of equipment, and lacks of generic applicability.

Images